Optimal production and maintenance scheduling for a degrading multi-failure modes single-machine production environment

• Published in: Applied soft computing Vol. 106; p. 107312
• Main Authors: Sharifi, Mani, Taghipour, Sharareh
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.07.2021
• Subjects: Deterioration; Genetic algorithm; Maintenance planning; Multi-state failure modes; Production scheduling; Simulated annealing; Single-machine; Teaching–learning-based​ optimization; Single-machine; Multi-state failure modes; Deterioration; Genetic algorithm; Teaching–learning-based​ optimization; Simulated annealing; Production scheduling; Maintenance planning
• ISSN: 1568-4946; 1872-9681
• DOI: 10.1016/j.asoc.2021.107312

Two important aspects of manufacturing systems are production scheduling and maintenance planning. These aspects are interdependent, but in most research work, this dependency is ignored. This paper proposed an integrated mathematical model for joint production scheduling and maintenance planning for a degrading multi-failure single-machine manufacturing system, in which the machine has discrete deterioration states. The machine is subject to different failure modes. The first failure mode is the machine’s full deterioration, which is detected at the end of a job’s processing, and the other failure mode is random breakdowns, which are detected at the time of failure. Two machine deterioration state-based thresholds are considered, and five different maintenance actions may carry out a replacement, preventive, and corrective perfect or imperfect maintenance. Since the machine’s states’ transitions follow an exponential distribution, a closed-form matrix-based mathematical model with probabilistic input parameters is presented. This paper aims to optimize the total system’s cost, including the maintenance cost, machine energy consumption cost as well as the makespan penalty for exceeding a pre-determined threshold. The proposed model determines the optimal jobs’ sequence as well as the machine’s deterioration state-based thresholds. Due to the complexity of the developed model, a genetic algorithm (GA), simulated annealing (SA) algorithm, and a teaching–learning-based optimization (TLBO) algorithm have been used to solve the presented model. The algorithms are validated using a full enumeration technique, and the model is validated by applying different maintenance strategies. Our results demonstrate the superiority of the GA compared to the other algorithms. •Presenting an integrated production scheduling and maintenance planning model.•Considering the machine discrete multi-deterioration states.•Considering two state-based thresholds for the machine maintenance actions.•Considering five different types of maintenance actions.•Presenting a closed-form matrix-based mathematical model with probabilistic input parameters.